System and method for the detection and termination of fraudulent services

Abstract

A system is disclosed for the automatic detection of fraudulent activity on a transaction network, for which each transaction over the network has an associated identifier. In one embodiment, the system includes voice comparison means for comparing a first sampled voice of a user of a first transaction with a subsequently sampled voice of a user of a subsequent transaction having an identical identifier to that of the first transaction. Control means in the form of a voice-based fraud detection engine is provided for determining, from said comparison, a profile of user usage that is representative of a total number of different users of the associated identifier. In a preferred embodiment, the system also includes voice sampling means for sampling a voice of the user of the first transaction to generate a first voice sample.

Description

RELATED APPLICATIONS [0001] This application is a continuation of and claims the benefit of PCT / AU2003 / 000978 under 35 U.S.C. §§ 365 and 120, filed on Jul. 31, 2003 and published in English on Feb. 5, 2004, which in turn takes priority from GB 0217769.9 and GB 0222643.9, each of which are incorporated by reference.FIELD OF THE INVENTION [0002] The present invention relates to a system and method for the detection and termination of fraudulent activity over a transaction network through the use of voice sampling means. In particular, but not exclusively, the invention relates to a method and system for detecting fraudulent activity implemented over a mobile or landline telecommunications network. BACKGROUND OF THE INVENTION [0003] Telecommunications networks permit the reliable, rapid, and secure exchange of information for individuals and businesses on local, regional, and global scales. Such networks are fundamental to the operation of all electronic commerce (e-commerce) and mobil...

AI Technical Summary

Benefits of technology

[0016] The profile of user usage provides an indication of the number of different users associated with a single transaction identifier. It is a particular benefit of the present invention that there is no need to positively identify the voice of the user from the voice sample(s), but rather a determination is made as to whether the number of negative identifications (i.e. the number of voice samples which do not correspond to a previously stored voice print) suggests fraudulent use. This reduces the required sophistication level of the voice comparison means forming part of the system. The present invention stems from the realisation by the inventor, that the number of different users there are of an identifier can be used to provide an indication of whether there is fraudulent use of the identifier. A further benefit of the invention is that it permits the use of high speed verification systems that generate small voice prints from voice samples, even though such systems may have relatively high error rates. As a result the throughput of the invention is maximised and its storage requirements are minimised.

[0017] In a preferred embodiment, the system also includes selection means configured to initiate sampling of the voice of the user by the voice sampling means only for selected transactions. This provides the advantage that the processing, voice sampling and data storage requirements of the system are reduced.

[0018] Preferably, the voice sampling means is configured to sample a voice of a user several times for each transaction. This provides the benefit that if the fraudster is aware of such voice sampling schemes, any attempt to circumvent the fraud detection process can be prevented.

[0020] This embodiment of the invention uses a combined fraud detection apparatus and voice verification technology to identify frauds, and offers important benefits to network operators and service providers. The invention achieves far greater reliability than conventional fraud detection systems, making it possible to automatically suspend services that are being exploited without the need for a fraud investigation team. As a result, two major cost savings are achieved. Firstly, the size of the fraud investigation team can be substantially reduced (or eliminated altogether) because it is no longer required to investigate every suspicious event that occurs on a network. Secondly, because the network operator does not have to wait for the fraudulent transaction to be confirmed by a manual investigation before it is stopped, the time required to stop each fraudulent transaction is reduced, limiting the loss incurred.

[0031] Preferably, the control means includes adjustment means for adjusting at least one of the thresholds for fraudulent use and the threshold for non-fraudulent use for an associated identifier in response to one or more of said associated identifier characteristics. In this way, different types of calls can have different thresholds associated therewith. For example, for calls originating from landline telephones, it may be expected that there will be a higher number of different users of the telephone than for a mobile telephone. The threshold beyond which there is an indication of fraudulent use can therefore be adjusted to further improve the success rate of the system.

[0033] In a preferred embodiment, the control means is arranged to automatically supply the fraud condition signal to the transaction network, thereby to implement automatic termination of the transaction in the event that the threshold for fraudulent use is exceeded. The fraud condition signal may also implement automatic termination of other services associated with the indentifier. This removes the need for human involvement in the analysis of fraud alerts generated by a fraud detection system, and thus reduces network operator costs associated with providing a fraud investigation team.

Problems solved by technology

The same problem is encountered with pre-paid call time.

The criminals are aided in this enterprise by the anonymity of telecommunications networks and the insecurity of their physical infrastructure.

Fraud is not only a serious problem for the operators of telecommunications networks, but also for the providers of services that make use of them.

In addition to the criminal acts of selling stolen services to third parties, the fraudulent use of credit and other payment cards for purchasing goods and services over telecommunications networks is a persistent occurrence.

One of the primary problems with such fraudulent acts is the cost implications of providing services that ultimately are not paid for.

However, systems such as this are relatively complex and, as a result, their reliability tends to be rather low.

The term identifier is used throughout this specification to refer to an identifying label of a network element or a facility or other thing used to conduct a transaction, which may be the subject of fraudulent use.

Images